- Email: [email protected]
Mean direct medical care costs associated with cervical cancer for commercially insured patients in Texas
YGYNO-976635; No. of pages: 6; 4C: Gynecologic Oncology xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Gynecologic Oncology journal homepage: www.elsevier.com/locate/ygyno
Mean direct medical care costs associated with cervical cancer for commercially insured patients in Texas David R Lairson a,⁎, Shuangshuang Fu b, Wenyaw Chan c, Li Xu d, Zeena Shelal e, Lois Ramondetta e a
Department of Management, Policy, and Community Health, School of Public Health, The University of Texas Health Science Center at Houston, 1200 Pressler St., Houston, TX 77030, USA Department of Epidemiology, Human Genetics & Environmental Sciences, School of Public Health, The University of Texas Health Science Center at Houston, 1200 Pressler St., Houston, TX 77030, USA c Department of Biostatistics, School of Public Health, The University of Texas Health Science Center at Houston, 1200 Pressler St., Houston, TX 77030, USA d Department of Head and Neck Surgery, Division of Surgery, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030, USA e Department of Gynecologic Oncology and Reproductive Medicine, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030, USA b
H I G H L I G H T S • • • •
Mean first year cost of new cervical cancer cases in Texas was $50,846. Mean second year cost of cervical cancer cases was $27,656. Cost declined steeply between month 1 and month 5 after diagnosis. Cost associated with co-morbidities and residing in west Texas.
a r t i c l e
i n f o
Article history: Received 12 December 2016 Received in revised form 3 February 2017 Accepted 6 February 2017 Available online xxxx Keywords: Health care costs Health expenditures Cervical cancer Cervix Cancer Claims Insurance
a b s t r a c t Objective. To determine the mean cervical cancer medical care costs for patients enrolled in commercial insurance in Texas. Cost is represented by insurer and patient payments for care. Methods. We estimated the mean medical care costs during the first 2 years after the index diagnosis date for patients with cervical cancer (cases). Cases were identified using claims-based International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9), diagnosis codes and matched to controls without a claimsbased ICD-9 code for cancer using a 2-step propensity score matching method. Index dates for the cases were randomly assigned to potential controls, and cases and controls were matched by index date. Data for cancer cases and controls were obtained from the de-identified 2011–2014 U.S. MarketScan databases. A generalized linear model was employed to compute the cost for censored months during the 2-year follow-up period. Differential costs were assessed by subtracting the medical costs incurred by controls from those incurred by cases. Results. During 2011–2014, 475 commercially insured Texas patients with newly diagnosed cervical cancer met the inclusion criteria. The first-year and second-year mean medical costs were $60,828 and $37,721 for cases and $9982 and $10,066 for controls, respectively. The differential costs of cervical cancer for the first and second years were $50,846 and $27,656, respectively. The major correlates of higher monthly cervical cancer costs were higher Charlson Comorbidity Index score during 6 months period prior to diagnosis, higher healthcare costs between 6 months and 3 months prior to diagnosis, and residence in the western region of Texas. Costs for cervical cancer patients decreased steeply between month 1 and month 5 after diagnosis and then were stable, while costs for the control group were stable throughout the follow-up period. Conclusions. Mean direct medical costs associated with cervical cancer in Texas were substantial. These data will serve as key cost parameters in models of costs associated with human papillomavirus (HPV)-related cancers in Texas and the economic evaluation of HPV vaccination dissemination in Texas. © 2017 Elsevier Inc. All rights reserved.
⁎ Corresponding author at: Division of Management, Policy, and Community Health, School of Public Health, The University of Texas Health Science Center at Houston, 1200 Pressler St., Houston, TX 77030, USA. E-mail addresses: [email protected] (D.R. Lairson), [email protected] (S. Fu), [email protected] (W. Chan), [email protected] (L. Xu), [email protected] (Z. Shelal), [email protected] (L. Ramondetta).
http://dx.doi.org/10.1016/j.ygyno.2017.02.011 0090-8258/© 2017 Elsevier Inc. All rights reserved.
Please cite this article as: D.R. Lairson, et al., Mean direct medical care costs associated with cervical cancer for commercially insured patients in Texas, Gynecol Oncol (2017), http://dx.doi.org/10.1016/j.ygyno.2017.02.011
2
D.R. Lairson et al. / Gynecologic Oncology xxx (2017) xxx–xxx
1. Introduction Human papillomavirus (HPV)-related cervical cancer continues to cause serious health and economic consequences despite the fact that both cervical cancer screening and HPV vaccination are available and covered by public and private insurance programs. From 2008 to 2012, the average annual number of new cases of HPV-associated cancers of the cervix in the United States was 11,771, and an estimated 90.6% of cases were attributable to HPV [1]. In Texas in 2015, there were about 1112 new cases of HPV-related cervical cancer, and over 390 deaths were attributed to the disease [2]. Texas is among the states with the highest cervical cancer incidence rates, reporting an age-adjusted incidence of 8.7 per 100,000 population in 2013, compared with a rate of 7.2 per 100,000 population in the US women. Cervical cancer age-adjusted death rate for Texas was 2.7 per 100,000 population, higher than the national rate of 2.3 per 100,000 population [3]. Of all HPV-related cancers, cervical cancer is the one associated with the highest direct treatment costs in the United States [4]: an estimated $441 million annually in 2010 U.S. dollars. Texas state Medicaid payments for all acute care for cervical cancer totaled $8.4 million in 2014 [5]. The HPV immunization rate is relatively low in Texas, with about 40% of girls and 24% of boys completing full vaccination series [6]. Given limited resources, it is important to consider the potential benefits of HPV vaccination initiatives designed to reduce cervical cancer in Texas. The leading economic decision analytic models of HPV immunization rely on earlier cost estimates for cervical cancer treatment [7–10]. Those earlier cost estimates are from the 1990s and are based on limited numbers of patients from a single health maintenance organization in the northwest United States; therefore, they are not generalizable to cervical cancer treatment in Texas today. The primary aim of this study was to estimate the mean first-year costs of treating new cases of cervical cancer in Texas. Cost is represented by insurer and patient payments for care. Additional aims were to estimate the mean 2-year cervical cancer treatment costs in Texas and examine the insurance, demographic, and comorbidity correlates of cervical cancer treatment costs in the state. In future work, these results can be used together with cost estimates for uninsured and publicly insured groups to model the expected total statewide costs of cervical cancer in Texas with and without increases in HPV immunization rates. 2. Methods 2.1. Data sources We identified Texas women with cervical cancer from the de-identified 2011–2014 U.S. MarketScan databases. The databases had between 160 and 206 million enrollees per year during the study period, of whom 92% were insured through a commercial plan and 8% were enrolled in the Medicare supplemental plan. Information on demographics, diagnosis, enrollment duration, and inpatient, outpatient, and pharmacy healthcare utilization and costs was extracted from the databases. 2.2. Study population Cervical cancer cases were identified from healthcare claims covering the period from January 1, 2011, to December 31, 2014. To qualify as an incident case, a woman had to 1) have either 1 inpatient claim or 2 outpatient claims at least 30 days apart with a primary or secondary diagnosis with an International Classification of Diseases, Ninth Revision (ICD-9), code for cervical cancer (180.0–180.9); 2) have been continuously enrolled for 6 months before and after the index diagnosis date, which was the first date when a cervical cancer diagnosis code appeared during the study period; and 3) be aged 18 years or older. We excluded cases with 1-year costs greater than $1 million U.S. dollars.
The control group was selected from the Texas female population without a claims-based ICD-9 code for HPV-related cancer or cancer at any site (140.0–208.9) and aged 18 or older. Two steps were used in the control selection process. In the first step, an initial group of population controls was selected on the basis of 4 matching criteria with respect to cancer cases: 1) index date identical to the index date of the case (index dates for cases were randomly assigned to all non-cases); 2) no cancer ICD-9 code during the 6 months prior to the index date, and 6 months of continuous enrollment before and after the index date; 3) age ± 5 years; and 4) insurance type (commercial vs. Medicare only). From this initial group, a single population control was then selected for each cancer case using nearest-available-Mahalanobis-metric matching within calipers defined by the propensity score [11,12]. The propensity score was produced from the following 5 covariates: 1) Charlson Comorbidity Index score during 6 months period before the index date (the Charlson Comorbidity Index was modified such that any malignancy, metastatic solid tumor, or chronic pulmonary disease were excluded) [13,14]; 2) count of the number of Psychiatric Diagnosis Groups during 6 months period before the index date [15]; 3) healthcare costs observed between 6 months and 3 months prior to the index date (costs incurred during the 3 months immediately prior to the index date were excluded because these costs were more likely to be associated with the cancer diagnostics and could cause biased estimates) [16]; 4) health plan type [basic/major medical; comprehensive; exclusive provider organization (EPO)/missing; health maintenance organization/point-of-service plan (POS) with capitation/POS; preferred provider organization (PPO); or consumer-directed health plan/high-deductible health plan (CDHP/HDHP)]; and 5) Texas region, based on 3-digit zip code area [northeast (750–752, 754–762, 764, and 766–767), southeast (765, 770, 773–787, and 789), and west (remaining zip codes)]. These regions include Dallas-Ft. Worth (northeast), Houston, Austin, Corpus Christi, and San Antonio (southeast/ central), and El Paso, Midland-Odessa, and Lubbock (west) and surrounding populations. These regions represent major geographic and cultural areas in Texas. 2.3. Economic outcome measures and statistical analysis Cost was measured as the total gross payment to a provider for a specific service (i.e., the amount eligible for payment after application of pricing guidelines, such as fee schedules and discounts, and before application of deductibles, copayments, and coordination of benefits). Overall, inpatient, outpatient, and pharmacy costs incurred during the 2 years after the index date were calculated and compared between case and control groups. Monthly costs in the first 2 years after the index date were also calculated. All costs were healthcare inflation adjusted to 2015 U.S. dollars using the medical care component of the United States Bureau of Labor Statistics Consumer Price Index [17]. For cases and controls with b 2 years of follow-up, a generalized linear model was employed to compute the costs for censored months using the log-link function for the cost variable [18]. Independent variables included in the model to predict the cost were age, Charlson Comorbidity Index score, number of Psychiatric Diagnosis Groups, health plan type, Texas region, case-control group indicator, costs incurred between 6 months and 3 months before prior to the index date, censor indicator, and a polynomial of months since the index date. Generalized polynomial regression was used for describing the non-linear relationship between month and cost. The model with the lowest Akaike information criterion value was chosen as the best fitted model to determine the degree of the polynomial [19]. Baseline characteristics for cases and controls were reported using numbers and percentages for categorical variables and means and standard deviations for continuous variables. The comparability of baseline characteristics between cases and controls was assessed using chisquare tests for categorical variables and t-tests for continuous variables.
Please cite this article as: D.R. Lairson, et al., Mean direct medical care costs associated with cervical cancer for commercially insured patients in Texas, Gynecol Oncol (2017), http://dx.doi.org/10.1016/j.ygyno.2017.02.011
D.R. Lairson et al. / Gynecologic Oncology xxx (2017) xxx–xxx
3. Results A total of 515 cervical cancer cases met the study eligibility criteria and were matched 1:1 to an equal number of population controls. Among the 515 cases, 475 had commercial insurance and were included in the analysis along with their matched population controls. The case and control selection processes are described in Supplemental Figs. 1 and 2. Baseline characteristics of these samples are described in Table 1. There were no substantial differences between cancer cases and population controls in any baseline characteristic except employment classification and years of follow-up. The mean monthly costs in years 1 and 2 after the index date were $5069 and $3143, respectively, for the case group and $832 and $839, respectively, for the control group. The mean annual costs in years 1 and 2 after the index date were $60,828 and $37,721, respectively, for the case group and $9982 and $10,066, respectively, for the control group. Therefore, the mean differential costs of cervical cancer versus control were $50,846 for the first year and $27,656 for the second year. Costs for the first year broken down by inpatient service, outpatient service, and pharmacy costs are shown in Table 2. The highest mean costs for both cases and controls were outpatient costs ($39,377 for cases vs. $6156 for controls), followed by inpatient costs ($16,065 for cases vs. $1517 for controls) and drug costs ($2148 for cases vs. $1547 for controls). Table 3 presents the estimated parameters from the general linear model regression. Significant correlates of higher monthly costs were higher Charlson Comorbidity Index score, higher costs between 6 months and 3 months prior to the index diagnosis date, and residence in the western region of Texas. Fig. 1 presents the monthly cost distribution for cases and controls over the 2-year period after the index date with costs predicted for censored months using the general linear model estimates. For cervical cancer cases, there was a steep drop in costs from the first month after cancer diagnosis through the fifth month. The cost remained level after the fifth month. In contrast, the cost for the control group was steady over the 2-year follow-up period. 4. Discussion The present study examined the mean annual and monthly differential costs associated with cervical cancer for a commercially insured population in Texas. We found that for cervical cancer patients identified by either primary or secondary diagnosis, higher direct medical care costs were incurred in the first 5 to 6 months after cancer diagnosis. Costs after the first 6 months were stable but still higher than costs for the control group. For the control group, mean yearly costs were similar to the mean yearly health expenditure of $9523 per person for the 2014 general U.S. population [20]. Mean yearly costs for the control group were higher than the mean yearly health expenditure of $7045 per person in Texas [21]. This finding was not surprising given that commercially insured enrollees have higher mean costs than the publicly insured population [22]. The regression analysis showed that costs were higher in the west region of Texas compared to the southeast and northeast regions. The west region includes the U.S. Mexico border area which includes the lowest income part of the State. Data from the Texas Cancer Registry from 1995 to 2013 shows that the West region has a higher percent of cases diagnosed at the regional stage (37.5) compared to the southeast (33.8) and the northeast (34.6) [23]. One-hundred and three of the 475 cervical cancer patients were identified by secondary diagnoses on the index date. There were no substantial differences between cases identified by primary vs. secondary diagnosis in any baseline characteristic except for the Charlson comorbidity score, Psychiatric Diagnosis Groups and cost between six months and three months prior to the index date (supplementary Table 1). These measures were slightly higher for cases identified by secondary diagnosis, indicating a potential for higher cost compared to cases identified by primary diagnosis. However, the differential cost between
3
Table 1 Baseline characteristics of cervical cancer patients and matched population controls.a Characteristic
Cervical cancer patients
Population controls
Total number of patients Age, mean (SD), y
475 46.60 (10.29) 1.55 (0.82)
475 46.54 (10.42) 1.71 (0.88)
331 (69.68) 144 (30.32)
353 (74.32) 122 (25.68)
62 (13.05) 0 (0.00) 15 (3.15) 52 (10.95) 24 (5.05) 27 (5.68) 32 (6.74) 6 (1.26) 257 (54.11)
88 (18.53) 1 (0.21) 30 (6.32) 64 (13.47) 36 (7.58) 18 (3.79) 21 (4.42) 3 (0.63) 214 (45.05)
205 (43.16) 3 (0.63) 23 (4.84) 7 (1.47) 2 (0.42) 2 (0.42) 1 (0.21) 1 (0.21) 231 (48.63)
284 (59.79) 7 (1.47) 16 (3.37) 5 (1.05) 3 (0.63) 0 (0.00) 1 (0.21) 2 (0.42) 157 (33.05)
9 (1.89) 18 (3.79) 64 (13.47)
10 (2.11) 13 (2.74) 75 (15.79)
346 (72.84) 38 (0.08)
335 (70.53) 42 (8.84)
249 (52.42) 184 (38.74) 42 (8.84) 0.38 (0.91)
262 (55.16) 175 (36.84) 38 (8.00) 0.36 (0.87)
0.687
0.15 (0.47)
0.15 (0.43)
0.943
3005 (8808)
2809 (12,514)
0.765
Follow-up time, mean (SD), y Follow-up time N1 year b1 year Employee classification Salary non-union Salary union Salary other Hourly non-union Hourly union Hourly other Non-union Union Unknown Employment status Active full time Active part time and seasonal Early retiree Medicare-eligible retiree Retiree COBRA Long-term disability Surviving spouse/dependent Other/unknown Health plan type Comprehensive EPO/missing HMO/POS with capitation/POS without capitation PPO CDHP/HDHP Region Southeast Northeast West Charlson Comorbidity Index score, mean (SD)b No. of Psychiatric Diagnosis Groups, mean (SD)b Costs between 6 months and 3 months prior to diagnosis, mean (SD), $c
P-value
0.928 0.003 0.112
0.003
b0.001
0.716
0.685
Abbreviations: CDHP, consumer-driven health plan; COBRA, Consolidated Omnibus Budget Reconciliation Act; EPO, exclusive provider organization; HDHP, high-deductible health plan; HMO, health maintenance organization; POS, point of service; PPO, preferred provider organization. a Values are expressed as number of patients (percentage) unless otherwise indicated. b Measured during 6 months period prior to the index date. c Costs incurred during the 3 months immediately prior to the index date were excluded to avoid including the costs of treating symptoms or diagnosing cancer.
cases and controls after the index date increased by b1% for year 1 and b10% for year 2 compared to only including cases with a primary diagnosis of cervical cancer (supplementary Table 2). Direct medical costs associated with cervical cancer were higher in our study than in studies previously published by other groups. Direct medical care costs for the Texas cervical cancer patients in our study were higher than direct medical costs for cervical cancer patients in a study conducted by Insinga et al., who analyzed the healthcare costs associated with cervical cancer in a large fee-for-service health plan population from 1998 through 2003 and estimated a differential 2-year cumulative cost per case of $16,608 (2015 U.S. dollars) associated with cervical cancer [24]. A study of cervical cancer costs in a Canadian population from 2007 through 2010 found, as in our present study, a decrease in costs between year 1 and year 2, but costs were lower than in our study: $34,704 in year 1 and $12,398 in year 2 (2015 U.S. dollars) [25]. Helms and Melnikow used micro-costing to estimate the costs of
Please cite this article as: D.R. Lairson, et al., Mean direct medical care costs associated with cervical cancer for commercially insured patients in Texas, Gynecol Oncol (2017), http://dx.doi.org/10.1016/j.ygyno.2017.02.011
4
D.R. Lairson et al. / Gynecologic Oncology xxx (2017) xxx–xxx
Table 2 Observed costs in U.S. dollars for the first year after the index date for cervical cancer patients and matched population controls. Cervical cancer patients (n = 475)
Matched population controls (n = 475)
Variable
Mean
SD
Median
Min
Max
Mean
SD
Median
Min
Max
All costs Inpatient service Outpatient service Drug
57,590 16,065 39,377 2148
76,959 34,040 56,028 7152
25,580 0 14,422 351
16 0 0 0
472,145 274,935 444,225 103,331
9221 1517 6156 1547
35,807 9272 33,635 4646
2216 0 1433 228
0 0 0 0
709,674 162,256 706,161 56,634
cervical cancer treatment with data from a health maintenance organization from 1990 through 1991, and they found a mean lifetime cost of treatment of $58,998 (2015 U.S. dollars) for cervical cancer [8]. A study by McCrory et al. using the 1992–1994 Medstat MarketScan database estimated the costs of cervical cancer for each stage as follows: stage I, $37,360; stage II-III, $57,313; and stage IV, 85,285 (2015 U.S. dollars) [26]. A number of factors may contribute to the differences in estimated costs of cervical cancer between previous studies and our current study. First, previous studies were based on cervical cancer patients treated in the 1990s. Differences in the diagnostic procedures and treatment regimens for cervical cancer could all affect the cost of healthcare. Second, previous studies were not focused on the Texas population. Historical data suggest that the mean healthcare expenditure per capita in Texas is lower than the mean healthcare expenditure per capita in the United States [27]. Third, healthcare costs for commercially insured younger patients with cervical cancer may be lower than healthcare costs for elderly patients with cervical cancer enrolled in Medicare because older populations are less likely to have cervical cancer screening [28], more likely to be diagnosed with late-stage cervical cancer [29,30], and more likely to die from the disease [30]. Our study provides evidence of the cost per case of cervical cancer in Texas, where only approximately 30% of eligible girls and women have received 3 doses of the HPV vaccine, a vaccination rate lower than the
U.S. mean [31]. Costs estimated in this study will serve as important current parameters in models assessing the potential costs and benefits of disseminating HPV vaccination more widely in Texas. This study has several strengths. First, we compared costs in cervical cancer patients with costs in a population control group matched on age, region of Texas, health plan type, baseline comorbidity, and intensity of healthcare costs prior to diagnosis, which allowed us to calculate the differential costs associated with cervical cancer. Second, to our knowledge, this study is the most up-to-date study of cervical cancer costs; previous studies were conducted in cervical cancer cohorts from N10 years ago. Third, our costs were estimated with data from a community-based population, not patients enrolled in clinical trials, in which many variables are controlled and the study population is selected; thus, our study is more likely than clinical trials-based studies to reflect the real costs of cervical cancer. Finally, the mean age of this cervical cancer group is similar to the mean age of cervical cancer patients in the United States according to Surveillance, Epidemiology, and End Results data for 2016 [32]. The present study also has limitations. Cancer cases were limited to the commercially insured population in Texas, and therefore our findings cannot be generalized to patients who are uninsured or covered by Medicare or Medicaid, who have been found to have lower costs per case for other HPV-related cancers [33]. The Texas Cancer Registry reports about 1000 new cases of cervical cancer per year [23]. Our
Table 3 Coefficients for generalized linear model, 2-year cost estimate. Parameter
Beta estimate
Standard error
95% confidence limits
Z
Pr N |Z|
Intercept Age Charlson Comorbidity Index scorea No. of Psychiatric Diagnosis Groupsa Health plan type EPO/missing HMO/POS PPO CDHP/HDHP Comprehensive Texas region Northeast West Southeast Case/control status Case Control Costs between 6 months and 3 months prior to diagnosisb Censor status Censored Non-censored Month Month2 Month3 Month4
7.1122 0.0044 0.1912 0.0755
0.3634 0.0046 0.0438 0.0792
6.3999 −0.0046 0.1053 −0.0798
7.8244 0.0134 0.2771 0.2309
19.57 0.96 4.36 0.95
b0.0001 0.3396 b0.0001⁎⁎,⁎⁎⁎ 0.3404
0.3125 0.0516 0.2430 0.3266 Ref
0.3163 0.2261 0.2094 0.2977
−0.3073 −0.3917 −0.1675 −0.2570
0.9323 0.4948 0.6534 0.9101
0.99 0.23 1.16 1.10
0.3231 0.8196 0.2460 0.2727
0.1777 0.3951 Ref
0.1027 0.1908
−0.0236 0.0210
0.3791 0.7692
1.73 2.07
0.0836 0.0384⁎
1.3610 Ref 1.60E-05
0.0950
1.1747
1.5472
14.32
b0.0001⁎⁎,⁎⁎⁎
2.30E-06
1.20E-05
2.10E-05
7.16
b0.0001⁎⁎,⁎⁎⁎
0.1079
−0.0648
0.3581
1.36
0.1739
0.0640 0.0136 0.0010 2.20E-05
−0.4363 0.0078 −0.0036 −1.60E-05
−0.1856 0.0611 0.0003 7.20E-05
−4.86 2.54 −1.67 1.25
b0.0001⁎⁎,⁎⁎⁎ 0.0112⁎
0.1467 Ref −0.3110 0.0345 −0.0016 2.80E-05
0.0948 0.2124
Abbreviations: CDHP, consumer-driven health plan; EPO, exclusive provider organization; HDHP, high-deductible health plan; HMO, health maintenance organization; POS, point of service; PPO, preferred provider organization; Ref, reference. a Measured 6 months prior to the index date. b Costs incurred during the 3 months immediately prior to the index date were excluded to avoid including the costs of treating symptoms or testing of an undiagnosed cancer. ⁎ Statistically significant at α = 0.05. ⁎⁎ Statistically significant at α = 0.01. ⁎⁎⁎ Statistically significant at α = 0.001.
Please cite this article as: D.R. Lairson, et al., Mean direct medical care costs associated with cervical cancer for commercially insured patients in Texas, Gynecol Oncol (2017), http://dx.doi.org/10.1016/j.ygyno.2017.02.011
D.R. Lairson et al. / Gynecologic Oncology xxx (2017) xxx–xxx
5
Fig. 1. Mean total medical care costs by month during the first 2 years after the index date.
study includes 475 cases over a four year period that was available in the private insurance data and met all of our inclusion criteria. We plan to conduct a similar analysis of mean cost for the Medicaid and Medicare populations in Texas. The estimates are based on payments by insurers and patients and therefore are not a true measure of resource costs. The estimates do not include indirect costs such as lost earnings due to morbidity and premature mortality, nor do they capture the pain and suffering incurred by patients and their families. Stages of cancer and death information were not available in the healthcare claims database. Lifetime costs of cancer are directly related to stage at diagnosis and costs in the terminal phase of cancer are high relative to the continuing care phase of treatment [34]. We had a relatively short follow-up time of 4 years and no information on stage at diagnosis and deaths. However, mean monthly costs were stable after 6 months from the index diagnosis date, suggesting that our short term cost estimates were not underestimated. HPV-related cervical cancer continues to generate serious health and economic consequences despite widespread screening and the availability of effective HPV vaccinations. With vaccination rates lagging in Texas and other states, state policy makers must use current data to assess the economic impact of public investments to boost HPV immunization rates. Current state-level cost estimates are required to identify the scope of the problem and for use as parameter values in decision analytic models for assessing multiple alternative policy strategies. Parameters in the current leading models are inadequate [35]. The upto-date results from the present study should be of value to Texas program managers and policymakers and to researchers interested in examining the economic consequences of cervical cancer in other jurisdictions.
5. Conclusion In conclusion, the present study found that in commercially insured patients in Texas with cervical cancer treated during 2011–2014, the first-year and second-year mean differential costs associated with cervical cancer were $50,846 and $27,656, respectively (2015 U.S. dollars).
Funding This study was supported by generous philanthropic contributions, including a contribution from the Lyda Hill Foundation to The University of Texas MD Anderson Cancer Center's Moon Shots Program. This research was accomplished within the Oropharynx Program at The University of Texas MD Anderson Cancer Center and funded in part through the Stiefel Oropharyngeal Research Fund.
Conflicts of interest The authors declare that there are no conflicts of interest. Author contributions DRL: Study concept and design; data collection, analysis, and interpretation; manuscript drafting and revision SF: Data collection, analysis, and interpretation; manuscript drafting and revision WC: Study concept and design; data interpretation; manuscript drafting and revision LX: Study concept and design; data interpretation; manuscript drafting and revision ZS: Data interpretation; manuscript drafting and revision LR: Study concept; data interpretation; manuscript drafting and revision Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ygyno.2017.02.011. References [1] L.J. Viens, S.J. Henley, M. Watson, L.E. Markowitz, C.C. Thomas, T.D. Thompson, et al., Human papillomavirus-associated cancers - United States, 2008–2012, MMWR Morb. Mortal. Wkly Rep. 65 (26) (2016 Jul 8) 661–666. [2] Texas Expected Numbers of Cancer Cases and Deaths [Internet]: Texas Cancer Registry; 2015 [updated June 29, 2015; cited 10/06/2016], 2015 (Available from: http:// www.dshs.texas.gov/tcr/statisticalData/2015expected/2015-Texas-Expected-Numbers-of-Cancer-Cases-and-Deaths.aspx). [3] U.S. Cancer Statistics Working Group, United States Cancer Statistics: 1999–2013 Incidence and Mortality Web-based Report, U.S. Department of Health and Human Services, Centers for Disease Control and Prevention and National Cancer Institute, Atlanta, 2016. [4] H.W. Chesson, D.U. Ekwueme, M. Saraiya, M. Watson, D.R. Lowy, L.E. Markowitz, Estimates of the annual direct medical costs of the prevention and treatment of disease associated with human papillomavirus in the United States, Vaccine 30 (42) (2012 Sep 14) 6016–6019. [5] Strategic Decision Support, Texas Health and Human Services Commission, Texas Medicaid Selected Cancer Sites State Fiscal Year 2014 Final. AHQP Claims Universe, 2016. [6] S. Reagan-Steiner, D. Yankey, J. Jeyarajah, L.D. Elam-Evans, C.R. Curtis, J. MacNeil, et al., National, regional, state, and selected local area vaccination coverage among adolescents aged 13–17 years - United States, 2015, MMWR Morb. Mortal. Wkly Rep. 65 (33) (Aug. 26 2016) 850–858. [7] J.J. Kim, N.G. Campos, S. Sy, E.A. Burger, J. Cuzick, P.E. Castle, et al., Inefficiencies and high-value improvements in U.S. cervical cancer screening practice: a cost-effectiveness analysis, Ann. Intern. Med. 163 (8) (Oct. 20 2015) 589–597. [8] L.J. Helms, J. Melnikow, Determining costs of health care services for cost-effectiveness analyses: the case of cervical cancer prevention and treatment, Med. Care 37 (7) (1999 Jul.) 652–661.
Please cite this article as: D.R. Lairson, et al., Mean direct medical care costs associated with cervical cancer for commercially insured patients in Texas, Gynecol Oncol (2017), http://dx.doi.org/10.1016/j.ygyno.2017.02.011
6
D.R. Lairson et al. / Gynecologic Oncology xxx (2017) xxx–xxx
[9] E.H. Elbasha, E.J. Dasbach, Impact of vaccinating boys and men against HPV in the United States, Vaccine 28 (42) (Oct. 4 2010) 6858–6867. [10] M. Brisson, J.F. Laprise, H.W. Chesson, M. Drolet, T. Malagon, M.C. Boily, et al., Health and economic impact of switching from a 4-valent to a 9-valent HPV vaccination program in the United States, J. Natl. Cancer Inst. 108 (1) (Oct. 4 2015)http://dx. doi.org/10.1093/jnci/djv282 (Print 2016 Jan). [11] R.B. D'Agostino Jr., Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group, Stat. Med. 17 (19) (Oct. 15 1998) 2265–2281. [12] D. Rubin, Matching to remove bias in observational studies: a review, Sankhya 35 (1973) 417–446 (Series A). [13] M.E. Charlson, P. Pompei, K.L. Ales, C.R. MacKenzie, A new method of classifying prognostic comorbidity in longitudinal studies: development and validation, J. Chronic Dis. 40 (5) (1987) 373–383. [14] W. D'Hoore, A. Bouckaert, C. Tilquin, Practical considerations on the use of the Charlson comorbidity index with administrative data bases, J. Clin. Epidemiol. 49 (12) (1996 Dec) 1429–1433. [15] M.L. Ashcraft, B.E. Fries, D.R. Nerenz, S.P. Falcon, S.V. Srivastava, C.Z. Lee, et al., A psychiatric patient classification system. An alternative to diagnosis-related groups, Med. Care 27 (5) (1989 May) 543–557. [16] L.E. Cipriano, D. Romanus, C.C. Earle, B.A. Neville, E.F. Halpern, G.S. Gazelle, et al., Lung cancer treatment costs, including patient responsibility, by disease stage and treatment modality, 1992 to 2003, Value Health 14 (1) (2011 Jan) 41–52. [17] Consumer Price Index-Medical Care. [Internet]: U.S. Department of Labor. Bureau of Labor Statistics. [cited 7-20-2016](Available from:) http://data.bls.gov/cgi-bin/ surveymost?cu. [18] H. Glick, J. Doshi, S. Sonnad, D. Polsky, Analyzing cost, Economic Evaluation in Clinical Trials, Oxford University Press, New York 2014, p. 96. [19] H. Akaike, Likelihood of a model and information criteria, J. Econ. 16 (1981) 3–14. [20] National Health Expenditure Fact Sheet [Internet]: Centers for Medicare & Medicaid Services, 2016 ([updated 08/10/2016; cited 8/2/2016]. Available from: https:// www.cms.gov/research-statistics-data-and-systems/statistics-trends-and-reports/ nationalhealthexpenddata/nhe-fact-sheet.html). [21] Health Care Expenditures Per Capita by State of Residence [Internet]: Kaiser Family Foundation, 2011 ([updated December 2011; cited 12/1/2016]. Available from: http://kff.org/other/state-indicator/health-spending-per-capita/?currentTimeframe= 0&selectedRows=%7B%22nested%22:%7B%22texas%22:%7B%7D%7D%7D). [22] L. Ku, M. Broaddus, Public and private health insurance: stacking up the costs, Health Aff. (Millwood). 27 (4) (2008 Jul–Aug) w318–w327.
[23] Texas Cancer Registry, Cancer Epidemiology and Surveillance Branch, Texas Department of State Health Services [Internet], Texas Cancer Registry, 2013 ([cited 11/04/2016]. Available from: http://www.dshs.state.tx.us/tcr/default.shtm). [24] R.P. Insinga, X. Ye, P.K. Singhal, G.W. Carides, Healthcare resource use and costs associated with cervical, vaginal and vulvar cancers in a large U.S. health plan, Gynecol. Oncol. 111 (2) (2008 Nov.) 188–196. [25] C. Pendrith, A. Thind, G.S. Zaric, S. Sarma, Costs of cervical cancer treatment: population-based estimates from Ontario, Curr. Oncol. 23 (2) (2016 Apr.) e109–e115. [26] D.C. McCrory, D.B. Matchar, L. Bastian, S. Datta, V. Hasselblad, J. Hickey, et al., Evaluation of cervical cytology, Evid. Rep. Technol. Assess. (Summ.) 5 (5) (1999 Jan.) 1–6. [27] Centers for Medicare & Medicaid Services, Office of the Actuary, National Health Statistics Group, National Health Expenditure Data: Health Expenditures by State of Residence, 2011. [28] R.P. Insinga, A.G. Glass, B.B. Rush, Pap screening in a U.S. health plan, Cancer Epidemiol. Biomark. Prev. 13 (3) (2004 Mar) 355–360. [29] J.M. Ferrante, E.C. Gonzalez, R.G. Roetzheim, N. Pal, L. Woodard, Clinical and demographic predictors of late-stage cervical cancer, Arch. Fam. Med. 9 (5) (2000 May) 439–445. [30] N. Howlader, A.M. Noone, M. Krapcho, D. Miller, K. Bishop, S.F. Altekruse, C.L. Kosary, M. Yu, J. Ruhl, Z. Tatalovich, A. Mariotto, D.R. Lewis, H.S. Chen, E.J. Feuer, K.A. Cronin (Eds.), SEER Cancer Statistics Review, 1975–2013, National Cancer Institute, 2016 (SEER web site). [31] Vaccination Coverage Levels in Texas and US – NIS-Teen [Internet]: Texas Department of State Health Services, 2015 ([updated December 4, 2015; cited 8/25/ 2016]. Available from: https://www.dshs.texas.gov/immunize/coverage/nis.shtm). [32] SEER Stat Fact Sheets: Cervix Uteri Cancer [Internet]: Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute [cited 8/25/ 2016](Available from:) http://seer.cancer.gov/statfacts/html/cervix.html. [33] J.J. Jacobson, J.B. Epstein, F.C. Eichmiller, T.B. Gibson, G.S. Carls, E. Vogtmann, et al., The cost burden of oral, oral pharyngeal, and salivary gland cancers in three groups: commercial insurance, Medicare, and Medicaid, Head Neck Oncol. 4 (15) (2012) (3284-4-15. Epub 2012 Apr 26). [34] M.L. Brown, G.F. Riley, N. Schussler, R. Etzioni, Estimating health care costs related to cancer treatment from SEER-Medicare data, Med. Care 40 (Suppl. 8) (2002 Aug) IV104–IV117. [35] Y. Jiang, A. Gauthier, M.J. Postma, L. Ribassin-Majed, N. Largeron, X. Bresse, A critical review of cost-effectiveness analyses of vaccinating males against human papillomavirus, Hum. Vaccin. Immunother. 9 (11) (2013 Nov) 2285–2295.
Please cite this article as: D.R. Lairson, et al., Mean direct medical care costs associated with cervical cancer for commercially insured patients in Texas, Gynecol Oncol (2017), http://dx.doi.org/10.1016/j.ygyno.2017.02.011
Contents lists available at ScienceDirect
Gynecologic Oncology journal homepage: www.elsevier.com/locate/ygyno
Mean direct medical care costs associated with cervical cancer for commercially insured patients in Texas David R Lairson a,⁎, Shuangshuang Fu b, Wenyaw Chan c, Li Xu d, Zeena Shelal e, Lois Ramondetta e a
Department of Management, Policy, and Community Health, School of Public Health, The University of Texas Health Science Center at Houston, 1200 Pressler St., Houston, TX 77030, USA Department of Epidemiology, Human Genetics & Environmental Sciences, School of Public Health, The University of Texas Health Science Center at Houston, 1200 Pressler St., Houston, TX 77030, USA c Department of Biostatistics, School of Public Health, The University of Texas Health Science Center at Houston, 1200 Pressler St., Houston, TX 77030, USA d Department of Head and Neck Surgery, Division of Surgery, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030, USA e Department of Gynecologic Oncology and Reproductive Medicine, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030, USA b
H I G H L I G H T S • • • •
Mean first year cost of new cervical cancer cases in Texas was $50,846. Mean second year cost of cervical cancer cases was $27,656. Cost declined steeply between month 1 and month 5 after diagnosis. Cost associated with co-morbidities and residing in west Texas.
a r t i c l e
i n f o
Article history: Received 12 December 2016 Received in revised form 3 February 2017 Accepted 6 February 2017 Available online xxxx Keywords: Health care costs Health expenditures Cervical cancer Cervix Cancer Claims Insurance
a b s t r a c t Objective. To determine the mean cervical cancer medical care costs for patients enrolled in commercial insurance in Texas. Cost is represented by insurer and patient payments for care. Methods. We estimated the mean medical care costs during the first 2 years after the index diagnosis date for patients with cervical cancer (cases). Cases were identified using claims-based International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9), diagnosis codes and matched to controls without a claimsbased ICD-9 code for cancer using a 2-step propensity score matching method. Index dates for the cases were randomly assigned to potential controls, and cases and controls were matched by index date. Data for cancer cases and controls were obtained from the de-identified 2011–2014 U.S. MarketScan databases. A generalized linear model was employed to compute the cost for censored months during the 2-year follow-up period. Differential costs were assessed by subtracting the medical costs incurred by controls from those incurred by cases. Results. During 2011–2014, 475 commercially insured Texas patients with newly diagnosed cervical cancer met the inclusion criteria. The first-year and second-year mean medical costs were $60,828 and $37,721 for cases and $9982 and $10,066 for controls, respectively. The differential costs of cervical cancer for the first and second years were $50,846 and $27,656, respectively. The major correlates of higher monthly cervical cancer costs were higher Charlson Comorbidity Index score during 6 months period prior to diagnosis, higher healthcare costs between 6 months and 3 months prior to diagnosis, and residence in the western region of Texas. Costs for cervical cancer patients decreased steeply between month 1 and month 5 after diagnosis and then were stable, while costs for the control group were stable throughout the follow-up period. Conclusions. Mean direct medical costs associated with cervical cancer in Texas were substantial. These data will serve as key cost parameters in models of costs associated with human papillomavirus (HPV)-related cancers in Texas and the economic evaluation of HPV vaccination dissemination in Texas. © 2017 Elsevier Inc. All rights reserved.
⁎ Corresponding author at: Division of Management, Policy, and Community Health, School of Public Health, The University of Texas Health Science Center at Houston, 1200 Pressler St., Houston, TX 77030, USA. E-mail addresses: [email protected] (D.R. Lairson), [email protected] (S. Fu), [email protected] (W. Chan), [email protected] (L. Xu), [email protected] (Z. Shelal), [email protected] (L. Ramondetta).
http://dx.doi.org/10.1016/j.ygyno.2017.02.011 0090-8258/© 2017 Elsevier Inc. All rights reserved.
Please cite this article as: D.R. Lairson, et al., Mean direct medical care costs associated with cervical cancer for commercially insured patients in Texas, Gynecol Oncol (2017), http://dx.doi.org/10.1016/j.ygyno.2017.02.011
2
D.R. Lairson et al. / Gynecologic Oncology xxx (2017) xxx–xxx
1. Introduction Human papillomavirus (HPV)-related cervical cancer continues to cause serious health and economic consequences despite the fact that both cervical cancer screening and HPV vaccination are available and covered by public and private insurance programs. From 2008 to 2012, the average annual number of new cases of HPV-associated cancers of the cervix in the United States was 11,771, and an estimated 90.6% of cases were attributable to HPV [1]. In Texas in 2015, there were about 1112 new cases of HPV-related cervical cancer, and over 390 deaths were attributed to the disease [2]. Texas is among the states with the highest cervical cancer incidence rates, reporting an age-adjusted incidence of 8.7 per 100,000 population in 2013, compared with a rate of 7.2 per 100,000 population in the US women. Cervical cancer age-adjusted death rate for Texas was 2.7 per 100,000 population, higher than the national rate of 2.3 per 100,000 population [3]. Of all HPV-related cancers, cervical cancer is the one associated with the highest direct treatment costs in the United States [4]: an estimated $441 million annually in 2010 U.S. dollars. Texas state Medicaid payments for all acute care for cervical cancer totaled $8.4 million in 2014 [5]. The HPV immunization rate is relatively low in Texas, with about 40% of girls and 24% of boys completing full vaccination series [6]. Given limited resources, it is important to consider the potential benefits of HPV vaccination initiatives designed to reduce cervical cancer in Texas. The leading economic decision analytic models of HPV immunization rely on earlier cost estimates for cervical cancer treatment [7–10]. Those earlier cost estimates are from the 1990s and are based on limited numbers of patients from a single health maintenance organization in the northwest United States; therefore, they are not generalizable to cervical cancer treatment in Texas today. The primary aim of this study was to estimate the mean first-year costs of treating new cases of cervical cancer in Texas. Cost is represented by insurer and patient payments for care. Additional aims were to estimate the mean 2-year cervical cancer treatment costs in Texas and examine the insurance, demographic, and comorbidity correlates of cervical cancer treatment costs in the state. In future work, these results can be used together with cost estimates for uninsured and publicly insured groups to model the expected total statewide costs of cervical cancer in Texas with and without increases in HPV immunization rates. 2. Methods 2.1. Data sources We identified Texas women with cervical cancer from the de-identified 2011–2014 U.S. MarketScan databases. The databases had between 160 and 206 million enrollees per year during the study period, of whom 92% were insured through a commercial plan and 8% were enrolled in the Medicare supplemental plan. Information on demographics, diagnosis, enrollment duration, and inpatient, outpatient, and pharmacy healthcare utilization and costs was extracted from the databases. 2.2. Study population Cervical cancer cases were identified from healthcare claims covering the period from January 1, 2011, to December 31, 2014. To qualify as an incident case, a woman had to 1) have either 1 inpatient claim or 2 outpatient claims at least 30 days apart with a primary or secondary diagnosis with an International Classification of Diseases, Ninth Revision (ICD-9), code for cervical cancer (180.0–180.9); 2) have been continuously enrolled for 6 months before and after the index diagnosis date, which was the first date when a cervical cancer diagnosis code appeared during the study period; and 3) be aged 18 years or older. We excluded cases with 1-year costs greater than $1 million U.S. dollars.
The control group was selected from the Texas female population without a claims-based ICD-9 code for HPV-related cancer or cancer at any site (140.0–208.9) and aged 18 or older. Two steps were used in the control selection process. In the first step, an initial group of population controls was selected on the basis of 4 matching criteria with respect to cancer cases: 1) index date identical to the index date of the case (index dates for cases were randomly assigned to all non-cases); 2) no cancer ICD-9 code during the 6 months prior to the index date, and 6 months of continuous enrollment before and after the index date; 3) age ± 5 years; and 4) insurance type (commercial vs. Medicare only). From this initial group, a single population control was then selected for each cancer case using nearest-available-Mahalanobis-metric matching within calipers defined by the propensity score [11,12]. The propensity score was produced from the following 5 covariates: 1) Charlson Comorbidity Index score during 6 months period before the index date (the Charlson Comorbidity Index was modified such that any malignancy, metastatic solid tumor, or chronic pulmonary disease were excluded) [13,14]; 2) count of the number of Psychiatric Diagnosis Groups during 6 months period before the index date [15]; 3) healthcare costs observed between 6 months and 3 months prior to the index date (costs incurred during the 3 months immediately prior to the index date were excluded because these costs were more likely to be associated with the cancer diagnostics and could cause biased estimates) [16]; 4) health plan type [basic/major medical; comprehensive; exclusive provider organization (EPO)/missing; health maintenance organization/point-of-service plan (POS) with capitation/POS; preferred provider organization (PPO); or consumer-directed health plan/high-deductible health plan (CDHP/HDHP)]; and 5) Texas region, based on 3-digit zip code area [northeast (750–752, 754–762, 764, and 766–767), southeast (765, 770, 773–787, and 789), and west (remaining zip codes)]. These regions include Dallas-Ft. Worth (northeast), Houston, Austin, Corpus Christi, and San Antonio (southeast/ central), and El Paso, Midland-Odessa, and Lubbock (west) and surrounding populations. These regions represent major geographic and cultural areas in Texas. 2.3. Economic outcome measures and statistical analysis Cost was measured as the total gross payment to a provider for a specific service (i.e., the amount eligible for payment after application of pricing guidelines, such as fee schedules and discounts, and before application of deductibles, copayments, and coordination of benefits). Overall, inpatient, outpatient, and pharmacy costs incurred during the 2 years after the index date were calculated and compared between case and control groups. Monthly costs in the first 2 years after the index date were also calculated. All costs were healthcare inflation adjusted to 2015 U.S. dollars using the medical care component of the United States Bureau of Labor Statistics Consumer Price Index [17]. For cases and controls with b 2 years of follow-up, a generalized linear model was employed to compute the costs for censored months using the log-link function for the cost variable [18]. Independent variables included in the model to predict the cost were age, Charlson Comorbidity Index score, number of Psychiatric Diagnosis Groups, health plan type, Texas region, case-control group indicator, costs incurred between 6 months and 3 months before prior to the index date, censor indicator, and a polynomial of months since the index date. Generalized polynomial regression was used for describing the non-linear relationship between month and cost. The model with the lowest Akaike information criterion value was chosen as the best fitted model to determine the degree of the polynomial [19]. Baseline characteristics for cases and controls were reported using numbers and percentages for categorical variables and means and standard deviations for continuous variables. The comparability of baseline characteristics between cases and controls was assessed using chisquare tests for categorical variables and t-tests for continuous variables.
Please cite this article as: D.R. Lairson, et al., Mean direct medical care costs associated with cervical cancer for commercially insured patients in Texas, Gynecol Oncol (2017), http://dx.doi.org/10.1016/j.ygyno.2017.02.011
D.R. Lairson et al. / Gynecologic Oncology xxx (2017) xxx–xxx
3. Results A total of 515 cervical cancer cases met the study eligibility criteria and were matched 1:1 to an equal number of population controls. Among the 515 cases, 475 had commercial insurance and were included in the analysis along with their matched population controls. The case and control selection processes are described in Supplemental Figs. 1 and 2. Baseline characteristics of these samples are described in Table 1. There were no substantial differences between cancer cases and population controls in any baseline characteristic except employment classification and years of follow-up. The mean monthly costs in years 1 and 2 after the index date were $5069 and $3143, respectively, for the case group and $832 and $839, respectively, for the control group. The mean annual costs in years 1 and 2 after the index date were $60,828 and $37,721, respectively, for the case group and $9982 and $10,066, respectively, for the control group. Therefore, the mean differential costs of cervical cancer versus control were $50,846 for the first year and $27,656 for the second year. Costs for the first year broken down by inpatient service, outpatient service, and pharmacy costs are shown in Table 2. The highest mean costs for both cases and controls were outpatient costs ($39,377 for cases vs. $6156 for controls), followed by inpatient costs ($16,065 for cases vs. $1517 for controls) and drug costs ($2148 for cases vs. $1547 for controls). Table 3 presents the estimated parameters from the general linear model regression. Significant correlates of higher monthly costs were higher Charlson Comorbidity Index score, higher costs between 6 months and 3 months prior to the index diagnosis date, and residence in the western region of Texas. Fig. 1 presents the monthly cost distribution for cases and controls over the 2-year period after the index date with costs predicted for censored months using the general linear model estimates. For cervical cancer cases, there was a steep drop in costs from the first month after cancer diagnosis through the fifth month. The cost remained level after the fifth month. In contrast, the cost for the control group was steady over the 2-year follow-up period. 4. Discussion The present study examined the mean annual and monthly differential costs associated with cervical cancer for a commercially insured population in Texas. We found that for cervical cancer patients identified by either primary or secondary diagnosis, higher direct medical care costs were incurred in the first 5 to 6 months after cancer diagnosis. Costs after the first 6 months were stable but still higher than costs for the control group. For the control group, mean yearly costs were similar to the mean yearly health expenditure of $9523 per person for the 2014 general U.S. population [20]. Mean yearly costs for the control group were higher than the mean yearly health expenditure of $7045 per person in Texas [21]. This finding was not surprising given that commercially insured enrollees have higher mean costs than the publicly insured population [22]. The regression analysis showed that costs were higher in the west region of Texas compared to the southeast and northeast regions. The west region includes the U.S. Mexico border area which includes the lowest income part of the State. Data from the Texas Cancer Registry from 1995 to 2013 shows that the West region has a higher percent of cases diagnosed at the regional stage (37.5) compared to the southeast (33.8) and the northeast (34.6) [23]. One-hundred and three of the 475 cervical cancer patients were identified by secondary diagnoses on the index date. There were no substantial differences between cases identified by primary vs. secondary diagnosis in any baseline characteristic except for the Charlson comorbidity score, Psychiatric Diagnosis Groups and cost between six months and three months prior to the index date (supplementary Table 1). These measures were slightly higher for cases identified by secondary diagnosis, indicating a potential for higher cost compared to cases identified by primary diagnosis. However, the differential cost between
3
Table 1 Baseline characteristics of cervical cancer patients and matched population controls.a Characteristic
Cervical cancer patients
Population controls
Total number of patients Age, mean (SD), y
475 46.60 (10.29) 1.55 (0.82)
475 46.54 (10.42) 1.71 (0.88)
331 (69.68) 144 (30.32)
353 (74.32) 122 (25.68)
62 (13.05) 0 (0.00) 15 (3.15) 52 (10.95) 24 (5.05) 27 (5.68) 32 (6.74) 6 (1.26) 257 (54.11)
88 (18.53) 1 (0.21) 30 (6.32) 64 (13.47) 36 (7.58) 18 (3.79) 21 (4.42) 3 (0.63) 214 (45.05)
205 (43.16) 3 (0.63) 23 (4.84) 7 (1.47) 2 (0.42) 2 (0.42) 1 (0.21) 1 (0.21) 231 (48.63)
284 (59.79) 7 (1.47) 16 (3.37) 5 (1.05) 3 (0.63) 0 (0.00) 1 (0.21) 2 (0.42) 157 (33.05)
9 (1.89) 18 (3.79) 64 (13.47)
10 (2.11) 13 (2.74) 75 (15.79)
346 (72.84) 38 (0.08)
335 (70.53) 42 (8.84)
249 (52.42) 184 (38.74) 42 (8.84) 0.38 (0.91)
262 (55.16) 175 (36.84) 38 (8.00) 0.36 (0.87)
0.687
0.15 (0.47)
0.15 (0.43)
0.943
3005 (8808)
2809 (12,514)
0.765
Follow-up time, mean (SD), y Follow-up time N1 year b1 year Employee classification Salary non-union Salary union Salary other Hourly non-union Hourly union Hourly other Non-union Union Unknown Employment status Active full time Active part time and seasonal Early retiree Medicare-eligible retiree Retiree COBRA Long-term disability Surviving spouse/dependent Other/unknown Health plan type Comprehensive EPO/missing HMO/POS with capitation/POS without capitation PPO CDHP/HDHP Region Southeast Northeast West Charlson Comorbidity Index score, mean (SD)b No. of Psychiatric Diagnosis Groups, mean (SD)b Costs between 6 months and 3 months prior to diagnosis, mean (SD), $c
P-value
0.928 0.003 0.112
0.003
b0.001
0.716
0.685
Abbreviations: CDHP, consumer-driven health plan; COBRA, Consolidated Omnibus Budget Reconciliation Act; EPO, exclusive provider organization; HDHP, high-deductible health plan; HMO, health maintenance organization; POS, point of service; PPO, preferred provider organization. a Values are expressed as number of patients (percentage) unless otherwise indicated. b Measured during 6 months period prior to the index date. c Costs incurred during the 3 months immediately prior to the index date were excluded to avoid including the costs of treating symptoms or diagnosing cancer.
cases and controls after the index date increased by b1% for year 1 and b10% for year 2 compared to only including cases with a primary diagnosis of cervical cancer (supplementary Table 2). Direct medical costs associated with cervical cancer were higher in our study than in studies previously published by other groups. Direct medical care costs for the Texas cervical cancer patients in our study were higher than direct medical costs for cervical cancer patients in a study conducted by Insinga et al., who analyzed the healthcare costs associated with cervical cancer in a large fee-for-service health plan population from 1998 through 2003 and estimated a differential 2-year cumulative cost per case of $16,608 (2015 U.S. dollars) associated with cervical cancer [24]. A study of cervical cancer costs in a Canadian population from 2007 through 2010 found, as in our present study, a decrease in costs between year 1 and year 2, but costs were lower than in our study: $34,704 in year 1 and $12,398 in year 2 (2015 U.S. dollars) [25]. Helms and Melnikow used micro-costing to estimate the costs of
Please cite this article as: D.R. Lairson, et al., Mean direct medical care costs associated with cervical cancer for commercially insured patients in Texas, Gynecol Oncol (2017), http://dx.doi.org/10.1016/j.ygyno.2017.02.011
4
D.R. Lairson et al. / Gynecologic Oncology xxx (2017) xxx–xxx
Table 2 Observed costs in U.S. dollars for the first year after the index date for cervical cancer patients and matched population controls. Cervical cancer patients (n = 475)
Matched population controls (n = 475)
Variable
Mean
SD
Median
Min
Max
Mean
SD
Median
Min
Max
All costs Inpatient service Outpatient service Drug
57,590 16,065 39,377 2148
76,959 34,040 56,028 7152
25,580 0 14,422 351
16 0 0 0
472,145 274,935 444,225 103,331
9221 1517 6156 1547
35,807 9272 33,635 4646
2216 0 1433 228
0 0 0 0
709,674 162,256 706,161 56,634
cervical cancer treatment with data from a health maintenance organization from 1990 through 1991, and they found a mean lifetime cost of treatment of $58,998 (2015 U.S. dollars) for cervical cancer [8]. A study by McCrory et al. using the 1992–1994 Medstat MarketScan database estimated the costs of cervical cancer for each stage as follows: stage I, $37,360; stage II-III, $57,313; and stage IV, 85,285 (2015 U.S. dollars) [26]. A number of factors may contribute to the differences in estimated costs of cervical cancer between previous studies and our current study. First, previous studies were based on cervical cancer patients treated in the 1990s. Differences in the diagnostic procedures and treatment regimens for cervical cancer could all affect the cost of healthcare. Second, previous studies were not focused on the Texas population. Historical data suggest that the mean healthcare expenditure per capita in Texas is lower than the mean healthcare expenditure per capita in the United States [27]. Third, healthcare costs for commercially insured younger patients with cervical cancer may be lower than healthcare costs for elderly patients with cervical cancer enrolled in Medicare because older populations are less likely to have cervical cancer screening [28], more likely to be diagnosed with late-stage cervical cancer [29,30], and more likely to die from the disease [30]. Our study provides evidence of the cost per case of cervical cancer in Texas, where only approximately 30% of eligible girls and women have received 3 doses of the HPV vaccine, a vaccination rate lower than the
U.S. mean [31]. Costs estimated in this study will serve as important current parameters in models assessing the potential costs and benefits of disseminating HPV vaccination more widely in Texas. This study has several strengths. First, we compared costs in cervical cancer patients with costs in a population control group matched on age, region of Texas, health plan type, baseline comorbidity, and intensity of healthcare costs prior to diagnosis, which allowed us to calculate the differential costs associated with cervical cancer. Second, to our knowledge, this study is the most up-to-date study of cervical cancer costs; previous studies were conducted in cervical cancer cohorts from N10 years ago. Third, our costs were estimated with data from a community-based population, not patients enrolled in clinical trials, in which many variables are controlled and the study population is selected; thus, our study is more likely than clinical trials-based studies to reflect the real costs of cervical cancer. Finally, the mean age of this cervical cancer group is similar to the mean age of cervical cancer patients in the United States according to Surveillance, Epidemiology, and End Results data for 2016 [32]. The present study also has limitations. Cancer cases were limited to the commercially insured population in Texas, and therefore our findings cannot be generalized to patients who are uninsured or covered by Medicare or Medicaid, who have been found to have lower costs per case for other HPV-related cancers [33]. The Texas Cancer Registry reports about 1000 new cases of cervical cancer per year [23]. Our
Table 3 Coefficients for generalized linear model, 2-year cost estimate. Parameter
Beta estimate
Standard error
95% confidence limits
Z
Pr N |Z|
Intercept Age Charlson Comorbidity Index scorea No. of Psychiatric Diagnosis Groupsa Health plan type EPO/missing HMO/POS PPO CDHP/HDHP Comprehensive Texas region Northeast West Southeast Case/control status Case Control Costs between 6 months and 3 months prior to diagnosisb Censor status Censored Non-censored Month Month2 Month3 Month4
7.1122 0.0044 0.1912 0.0755
0.3634 0.0046 0.0438 0.0792
6.3999 −0.0046 0.1053 −0.0798
7.8244 0.0134 0.2771 0.2309
19.57 0.96 4.36 0.95
b0.0001 0.3396 b0.0001⁎⁎,⁎⁎⁎ 0.3404
0.3125 0.0516 0.2430 0.3266 Ref
0.3163 0.2261 0.2094 0.2977
−0.3073 −0.3917 −0.1675 −0.2570
0.9323 0.4948 0.6534 0.9101
0.99 0.23 1.16 1.10
0.3231 0.8196 0.2460 0.2727
0.1777 0.3951 Ref
0.1027 0.1908
−0.0236 0.0210
0.3791 0.7692
1.73 2.07
0.0836 0.0384⁎
1.3610 Ref 1.60E-05
0.0950
1.1747
1.5472
14.32
b0.0001⁎⁎,⁎⁎⁎
2.30E-06
1.20E-05
2.10E-05
7.16
b0.0001⁎⁎,⁎⁎⁎
0.1079
−0.0648
0.3581
1.36
0.1739
0.0640 0.0136 0.0010 2.20E-05
−0.4363 0.0078 −0.0036 −1.60E-05
−0.1856 0.0611 0.0003 7.20E-05
−4.86 2.54 −1.67 1.25
b0.0001⁎⁎,⁎⁎⁎ 0.0112⁎
0.1467 Ref −0.3110 0.0345 −0.0016 2.80E-05
0.0948 0.2124
Abbreviations: CDHP, consumer-driven health plan; EPO, exclusive provider organization; HDHP, high-deductible health plan; HMO, health maintenance organization; POS, point of service; PPO, preferred provider organization; Ref, reference. a Measured 6 months prior to the index date. b Costs incurred during the 3 months immediately prior to the index date were excluded to avoid including the costs of treating symptoms or testing of an undiagnosed cancer. ⁎ Statistically significant at α = 0.05. ⁎⁎ Statistically significant at α = 0.01. ⁎⁎⁎ Statistically significant at α = 0.001.
Please cite this article as: D.R. Lairson, et al., Mean direct medical care costs associated with cervical cancer for commercially insured patients in Texas, Gynecol Oncol (2017), http://dx.doi.org/10.1016/j.ygyno.2017.02.011
D.R. Lairson et al. / Gynecologic Oncology xxx (2017) xxx–xxx
5
Fig. 1. Mean total medical care costs by month during the first 2 years after the index date.
study includes 475 cases over a four year period that was available in the private insurance data and met all of our inclusion criteria. We plan to conduct a similar analysis of mean cost for the Medicaid and Medicare populations in Texas. The estimates are based on payments by insurers and patients and therefore are not a true measure of resource costs. The estimates do not include indirect costs such as lost earnings due to morbidity and premature mortality, nor do they capture the pain and suffering incurred by patients and their families. Stages of cancer and death information were not available in the healthcare claims database. Lifetime costs of cancer are directly related to stage at diagnosis and costs in the terminal phase of cancer are high relative to the continuing care phase of treatment [34]. We had a relatively short follow-up time of 4 years and no information on stage at diagnosis and deaths. However, mean monthly costs were stable after 6 months from the index diagnosis date, suggesting that our short term cost estimates were not underestimated. HPV-related cervical cancer continues to generate serious health and economic consequences despite widespread screening and the availability of effective HPV vaccinations. With vaccination rates lagging in Texas and other states, state policy makers must use current data to assess the economic impact of public investments to boost HPV immunization rates. Current state-level cost estimates are required to identify the scope of the problem and for use as parameter values in decision analytic models for assessing multiple alternative policy strategies. Parameters in the current leading models are inadequate [35]. The upto-date results from the present study should be of value to Texas program managers and policymakers and to researchers interested in examining the economic consequences of cervical cancer in other jurisdictions.
5. Conclusion In conclusion, the present study found that in commercially insured patients in Texas with cervical cancer treated during 2011–2014, the first-year and second-year mean differential costs associated with cervical cancer were $50,846 and $27,656, respectively (2015 U.S. dollars).
Funding This study was supported by generous philanthropic contributions, including a contribution from the Lyda Hill Foundation to The University of Texas MD Anderson Cancer Center's Moon Shots Program. This research was accomplished within the Oropharynx Program at The University of Texas MD Anderson Cancer Center and funded in part through the Stiefel Oropharyngeal Research Fund.
Conflicts of interest The authors declare that there are no conflicts of interest. Author contributions DRL: Study concept and design; data collection, analysis, and interpretation; manuscript drafting and revision SF: Data collection, analysis, and interpretation; manuscript drafting and revision WC: Study concept and design; data interpretation; manuscript drafting and revision LX: Study concept and design; data interpretation; manuscript drafting and revision ZS: Data interpretation; manuscript drafting and revision LR: Study concept; data interpretation; manuscript drafting and revision Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ygyno.2017.02.011. References [1] L.J. Viens, S.J. Henley, M. Watson, L.E. Markowitz, C.C. Thomas, T.D. Thompson, et al., Human papillomavirus-associated cancers - United States, 2008–2012, MMWR Morb. Mortal. Wkly Rep. 65 (26) (2016 Jul 8) 661–666. [2] Texas Expected Numbers of Cancer Cases and Deaths [Internet]: Texas Cancer Registry; 2015 [updated June 29, 2015; cited 10/06/2016], 2015 (Available from: http:// www.dshs.texas.gov/tcr/statisticalData/2015expected/2015-Texas-Expected-Numbers-of-Cancer-Cases-and-Deaths.aspx). [3] U.S. Cancer Statistics Working Group, United States Cancer Statistics: 1999–2013 Incidence and Mortality Web-based Report, U.S. Department of Health and Human Services, Centers for Disease Control and Prevention and National Cancer Institute, Atlanta, 2016. [4] H.W. Chesson, D.U. Ekwueme, M. Saraiya, M. Watson, D.R. Lowy, L.E. Markowitz, Estimates of the annual direct medical costs of the prevention and treatment of disease associated with human papillomavirus in the United States, Vaccine 30 (42) (2012 Sep 14) 6016–6019. [5] Strategic Decision Support, Texas Health and Human Services Commission, Texas Medicaid Selected Cancer Sites State Fiscal Year 2014 Final. AHQP Claims Universe, 2016. [6] S. Reagan-Steiner, D. Yankey, J. Jeyarajah, L.D. Elam-Evans, C.R. Curtis, J. MacNeil, et al., National, regional, state, and selected local area vaccination coverage among adolescents aged 13–17 years - United States, 2015, MMWR Morb. Mortal. Wkly Rep. 65 (33) (Aug. 26 2016) 850–858. [7] J.J. Kim, N.G. Campos, S. Sy, E.A. Burger, J. Cuzick, P.E. Castle, et al., Inefficiencies and high-value improvements in U.S. cervical cancer screening practice: a cost-effectiveness analysis, Ann. Intern. Med. 163 (8) (Oct. 20 2015) 589–597. [8] L.J. Helms, J. Melnikow, Determining costs of health care services for cost-effectiveness analyses: the case of cervical cancer prevention and treatment, Med. Care 37 (7) (1999 Jul.) 652–661.
Please cite this article as: D.R. Lairson, et al., Mean direct medical care costs associated with cervical cancer for commercially insured patients in Texas, Gynecol Oncol (2017), http://dx.doi.org/10.1016/j.ygyno.2017.02.011
6
D.R. Lairson et al. / Gynecologic Oncology xxx (2017) xxx–xxx
[9] E.H. Elbasha, E.J. Dasbach, Impact of vaccinating boys and men against HPV in the United States, Vaccine 28 (42) (Oct. 4 2010) 6858–6867. [10] M. Brisson, J.F. Laprise, H.W. Chesson, M. Drolet, T. Malagon, M.C. Boily, et al., Health and economic impact of switching from a 4-valent to a 9-valent HPV vaccination program in the United States, J. Natl. Cancer Inst. 108 (1) (Oct. 4 2015)http://dx. doi.org/10.1093/jnci/djv282 (Print 2016 Jan). [11] R.B. D'Agostino Jr., Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group, Stat. Med. 17 (19) (Oct. 15 1998) 2265–2281. [12] D. Rubin, Matching to remove bias in observational studies: a review, Sankhya 35 (1973) 417–446 (Series A). [13] M.E. Charlson, P. Pompei, K.L. Ales, C.R. MacKenzie, A new method of classifying prognostic comorbidity in longitudinal studies: development and validation, J. Chronic Dis. 40 (5) (1987) 373–383. [14] W. D'Hoore, A. Bouckaert, C. Tilquin, Practical considerations on the use of the Charlson comorbidity index with administrative data bases, J. Clin. Epidemiol. 49 (12) (1996 Dec) 1429–1433. [15] M.L. Ashcraft, B.E. Fries, D.R. Nerenz, S.P. Falcon, S.V. Srivastava, C.Z. Lee, et al., A psychiatric patient classification system. An alternative to diagnosis-related groups, Med. Care 27 (5) (1989 May) 543–557. [16] L.E. Cipriano, D. Romanus, C.C. Earle, B.A. Neville, E.F. Halpern, G.S. Gazelle, et al., Lung cancer treatment costs, including patient responsibility, by disease stage and treatment modality, 1992 to 2003, Value Health 14 (1) (2011 Jan) 41–52. [17] Consumer Price Index-Medical Care. [Internet]: U.S. Department of Labor. Bureau of Labor Statistics. [cited 7-20-2016](Available from:) http://data.bls.gov/cgi-bin/ surveymost?cu. [18] H. Glick, J. Doshi, S. Sonnad, D. Polsky, Analyzing cost, Economic Evaluation in Clinical Trials, Oxford University Press, New York 2014, p. 96. [19] H. Akaike, Likelihood of a model and information criteria, J. Econ. 16 (1981) 3–14. [20] National Health Expenditure Fact Sheet [Internet]: Centers for Medicare & Medicaid Services, 2016 ([updated 08/10/2016; cited 8/2/2016]. Available from: https:// www.cms.gov/research-statistics-data-and-systems/statistics-trends-and-reports/ nationalhealthexpenddata/nhe-fact-sheet.html). [21] Health Care Expenditures Per Capita by State of Residence [Internet]: Kaiser Family Foundation, 2011 ([updated December 2011; cited 12/1/2016]. Available from: http://kff.org/other/state-indicator/health-spending-per-capita/?currentTimeframe= 0&selectedRows=%7B%22nested%22:%7B%22texas%22:%7B%7D%7D%7D). [22] L. Ku, M. Broaddus, Public and private health insurance: stacking up the costs, Health Aff. (Millwood). 27 (4) (2008 Jul–Aug) w318–w327.
[23] Texas Cancer Registry, Cancer Epidemiology and Surveillance Branch, Texas Department of State Health Services [Internet], Texas Cancer Registry, 2013 ([cited 11/04/2016]. Available from: http://www.dshs.state.tx.us/tcr/default.shtm). [24] R.P. Insinga, X. Ye, P.K. Singhal, G.W. Carides, Healthcare resource use and costs associated with cervical, vaginal and vulvar cancers in a large U.S. health plan, Gynecol. Oncol. 111 (2) (2008 Nov.) 188–196. [25] C. Pendrith, A. Thind, G.S. Zaric, S. Sarma, Costs of cervical cancer treatment: population-based estimates from Ontario, Curr. Oncol. 23 (2) (2016 Apr.) e109–e115. [26] D.C. McCrory, D.B. Matchar, L. Bastian, S. Datta, V. Hasselblad, J. Hickey, et al., Evaluation of cervical cytology, Evid. Rep. Technol. Assess. (Summ.) 5 (5) (1999 Jan.) 1–6. [27] Centers for Medicare & Medicaid Services, Office of the Actuary, National Health Statistics Group, National Health Expenditure Data: Health Expenditures by State of Residence, 2011. [28] R.P. Insinga, A.G. Glass, B.B. Rush, Pap screening in a U.S. health plan, Cancer Epidemiol. Biomark. Prev. 13 (3) (2004 Mar) 355–360. [29] J.M. Ferrante, E.C. Gonzalez, R.G. Roetzheim, N. Pal, L. Woodard, Clinical and demographic predictors of late-stage cervical cancer, Arch. Fam. Med. 9 (5) (2000 May) 439–445. [30] N. Howlader, A.M. Noone, M. Krapcho, D. Miller, K. Bishop, S.F. Altekruse, C.L. Kosary, M. Yu, J. Ruhl, Z. Tatalovich, A. Mariotto, D.R. Lewis, H.S. Chen, E.J. Feuer, K.A. Cronin (Eds.), SEER Cancer Statistics Review, 1975–2013, National Cancer Institute, 2016 (SEER web site). [31] Vaccination Coverage Levels in Texas and US – NIS-Teen [Internet]: Texas Department of State Health Services, 2015 ([updated December 4, 2015; cited 8/25/ 2016]. Available from: https://www.dshs.texas.gov/immunize/coverage/nis.shtm). [32] SEER Stat Fact Sheets: Cervix Uteri Cancer [Internet]: Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute [cited 8/25/ 2016](Available from:) http://seer.cancer.gov/statfacts/html/cervix.html. [33] J.J. Jacobson, J.B. Epstein, F.C. Eichmiller, T.B. Gibson, G.S. Carls, E. Vogtmann, et al., The cost burden of oral, oral pharyngeal, and salivary gland cancers in three groups: commercial insurance, Medicare, and Medicaid, Head Neck Oncol. 4 (15) (2012) (3284-4-15. Epub 2012 Apr 26). [34] M.L. Brown, G.F. Riley, N. Schussler, R. Etzioni, Estimating health care costs related to cancer treatment from SEER-Medicare data, Med. Care 40 (Suppl. 8) (2002 Aug) IV104–IV117. [35] Y. Jiang, A. Gauthier, M.J. Postma, L. Ribassin-Majed, N. Largeron, X. Bresse, A critical review of cost-effectiveness analyses of vaccinating males against human papillomavirus, Hum. Vaccin. Immunother. 9 (11) (2013 Nov) 2285–2295.
Please cite this article as: D.R. Lairson, et al., Mean direct medical care costs associated with cervical cancer for commercially insured patients in Texas, Gynecol Oncol (2017), http://dx.doi.org/10.1016/j.ygyno.2017.02.011